While proteins produced in prokaryotes, for example Escherichia coli, have no carbohydrate chain, proteins and lipids produced in eukaryotes, such as yeast, fungi, plant cells and animal cells, have a carbohydrate chain bound thereto in many instances.
Carbohydrate chains bound to proteins in animal cells include N-glycoside bond type carbohydrate chains (also called N-glycans) bound to an asparagine (Asn) residue in the protein and O-glycoside bond type carbohydrate chains (also called O-glycans) bound to a serine (Ser) or threonine (Thr) residue. It has recently been revealed that a certain kind of lipid containing a carbohydrate chain is covalently bound to a number of proteins and that those proteins are attached to the cell membrane through the lipid. This carbohydrate chain-containing lipid is called glycosyl phosphatidylinositol anchor.
Other carbohydrate chains, including glycosaminoglycans, are also present in animal cells. Compounds comprising a protein covalently bound to a glycosaminoglycan are called proteoglycans. The glycosaminoglycans constituting the carbohydrate chains of proteoglycans are similar in structure to O-glycans, which are carbohydrate chains of glycoproteins, but differ chemically therefrom. Glycosaminoglycans comprise repeating disaccharide units composed of glucosamine or galactosamine and a uronic acid (except for keratan sulfate which has no uronic acid residue) having sulfate residues covalently bound thereto (except for hyaluronic acid which has no sulfate residue).
Further carbohydrate chains in animal cells are present in substances called glycolipids. Sphingoglycolipids are one type of glycolipid present in animal cells. Sphingoglycolipids are composed of a carbohydrate, a long-chain fatty acid and sphingosine, a long-chain base, covalently bound together. Glyceroglycolipids are composed of a carbohydrate chain and glycerol covalently bound together.
Recent advances in molecular biology and cellular biology have made it possible to clarify the functions of carbohydrate chains. To date, a variety of functions of carbohydrate chains have been elucidated. First, carbohydrate chains play an important role in the clearance of glycoproteins in blood. It is known that erythropoietin produced by introducing the relevant gene into Escherichia coli retains activity in vitro but undergoes rapid clearance in vivo [Dordal et al.: Endocrinology, 116, 2293 (1985) and Browne et al.: Cold Spring Harbor Symposia on Quantitative Biology, 51, 693 (1986)]. It is known that while native human granulocyte-macrophage colony stimulating factor (hGM-CSF) has two carbohydrate chains of the N-glycoside bond type, a reduction in the number of carbohydrate chains results in a proportional increase in the rate of clearance from rat plasma [Donahue et al.: Cold Spring Harbor Symposia on Quantitative Biology, 51, 685 (1986)]. The rate of clearance and the site of clearance may vary or differ depending on the structure of the carbohydrate chain in question. Thus, it is known that hGM-CSF having a sialic acid residue undergoes clearance in the kidney while hGM-CSF lacking sialic acid has an increased rate of clearance and undergoes clearance in the liver. Alpha1-acid glycoproteins differing in carbohydrate structure and biosynthesized in the presence of various N-glycoside type carbohydrate chain biosynthesis inhibitors using a rat liver primary culture system were studied with respect to their rate of clearance from rat plasma and their rate of clearance from rat perfusate. In both cases, the rate of clearance was reduced in the order: high mannose type, carbohydrate chain-deficient type, hybrid type and composite type (natural type). It is known that the clearance from blood of tissue-type plasminogen activator (t-PA), which is used as a thrombolytic agent, is greatly influenced by the structure of its carbohydrate chain.
It is known that carbohydrate chains give protease resistance to proteins. For example, when the carbohydrate formation on fibronectin is inhibited with tunicamycin, the rate of degradation of intracellular carbohydrate chain-deficient fibronectin increases. It is also known that addition of a carbohydrate chain may result in increased heat stability or freezing resistance. In the case of erythropoietin and β-interferon, among others, the carbohydrate chain is known to contribute to increased solubility of the protein.
Carbohydrate chains also serve to maintain protein tertiary structure. It is known that when the membrane binding protein of vesicular stomatitis virus is devoid of the two naturally-occurring N-glycoside bond type carbohydrate chains, transport of the protein to the cell surface is inhibited and that when new carbohydrate chains are added to the protein, it is transported. It was revealed that, in that case, intermolecular association of the protein through disulfide bonding is induced following the elimination of carbohydrate chains and, as a result, protein transport is inhibited. When carbohydrate chains are added, the association is inhibited and the proper tertiary protein structure is maintained and protein transport becomes possible. As regards the site of addition of the new carbohydrate, it has been shown that there is a considerable amount of flexibility. In contrast, it has also been shown in certain instances that, depending on the site of carbohydrate chain introduction, the transport of a protein having a natural carbohydrate chain or chains may be completely inhibited.
Examples are also known where a carbohydrate chain serves to mask an antigenic site of a polypeptide. In the case of hGM-CSF, prolactin, interferon-γ, Rauscher leukemia virus gp70 and influenza hemagglutinin, experiments using a polyclonal antibody or a monoclonal antibody directed to a specific site on the peptide suggest that carbohydrate chains of these proteins inhibit antibody binding. Cases are also known where carbohydrate chains themselves are directly involved in the expression of activity by a glycoprotein. For instance, carbohydrates are thought to be associated with the expression of activity of such glycoprotein hormones as luteinizing hormone, follicle stimulating hormone and chorionic gonadotropin.
Carbohydrate chains serve an important function in the phenomenon of recognition between cells, between proteins or between a cell and a protein. For example, it is known that structurally different carbohydrate chains undergo clearance in vivo at different sites. It has recently been revealed that the ligand of the protein ELAM-1 (also called E-selectin), which is expressed specifically on vascular endothelial cells during an inflammatory response and promotes adhesion to neutrophils, is a carbohydrate chain called sialyl Lewis x [NeuAcα2-3Galβ1-4(Fucα1-3)GlcNAc; where NeuAc: sialic acid; Gal: galactose; Fuc: fucose; GlcNAc: N-acetylglucosamine]. The possible use of carbohydrate chains themselves or modifications thereof as drugs or the like is thus suggested [Phillips et al.: Science, 250, 1130 (1990); Goelz et al.: Trends in Glyco-science and Glycotechnology, 4, 14-24 (1992)]. Like ELAM-1, L-selectin, expressed in some T lymphocytes or neutrophils, and GMP-140 (also called P-selectin), expressed in platelets or on the membrane surface of vascular endothelial cells upon inflammatory stimulation, are associated with inflammatory responses. It is suggested that their ligand may be a carbohydrate chain analogous to sialyl Lewis x, the ELAM-1 ligand [Rosen et al.: Trends in Glycoscience and Glycotechnology, 4, 1-13 (1992); Larsen et al.: Trends in Glycoscience and Glycotechnology, 4, 25-31 (1992); Aruffo et al.: Trends in Glycoscience and Glycotechnology, 4, 146-151 (1992)]. ELAM-1, GMP-140 and L-selectin are structurally similar to one another and are collectively called selectins.
It has been suggested that, in cancer metastatis, as in inflammatory responses, ELAM-1 and GMP-140 cause adhesion of cancer cells to the vascular endothelium or aggregation of cancer cells with platelets and thereby promote cancer metastatis [Goelz et al.: Trends in Glycoscience and Glycotechnology, 4, 14-24 (1992); Larsen et al.: Trends in Glycoscience and Glycotechnology, 4, 25-31 (1992)]. This is in agreement with the finding that the level of expression of the sialyl Lewis x carbohydrate chain is high in cancer cells that are highly metastatic [Irimura et al.: Jikken Igaku (Experimental Medicine), 6, 33-39 (1988)]. ELAM-1 binds not only to sialyl Lewis x but also to a carbohydrate chain called sialyl Lewis a [NeuAcα2-3Galβ1-3(Fucα1-4)GlcNAc]. The binding affinity for sialyl Lewis a is somewhat stronger [Berg et al.: Journal of Biological Chemistry, 266, 14869-14872 (1991); Takada et al.: Biochemical and Biophysical Research Communications, 179, 713-719 (1991); Larkin et al.: Journal of Biological Chemistry, 267, 13661-13668 (1992)]. The sialyl Lewis a carbohydrate chain is also a carbohydrate chain antigen expressed upon oncogenesis of cells and is reportedly correlated with cancer metastatis [Kannagi and Takada: Jikken Igaku (Experimental Medicine), 10, 96-107 (1992)].
Based on these findings, it is expected that the sialyl Lewis x carbohydrate chain and sialyl Lewis a carbohydrate chain or derivatives thereof might produce a strong anti-inflammatory effect through their binding to ELAM-1, L-selectin or GMP-140 and, further, might inhibit cancer metastatis.
In view of the above-mentioned mechanisms of inflammatory responses and cancer metastatis, it may be possible to suppress inflammatory responses or prevent cancer metastatis by suppressing the expression of glycosyltransferases which control the synthesis of ligand carbohydrate chains recognizable by ELAM-1, L-selectin or GMP-140. The antisense RNA/antisense DNA technique [Tokuhisa: Bioscience and Industry, 50, 322-326 (1992); Murakami: Kagaku (Chemistry), 46, 681-684 (1991)] and the triple helix technique [Chubb and Hogan: Trends in Biotechnology, 10, 132-136 (1992)] are useful in suppressing the expression of a certain specific gene. For suppressing the expression of a specific glycosyltransferase using the antisense RNA/DNA technique, information is necessary about the gene or the base sequence of the gene and therefore it is important to clone the gene encoding the specific glycosyltransferase and determine the base sequence of same.
It is possible to diagnose an inflammatory disease by detecting expression of a specific glycosyltransferase in inflammatory leukocytes. In addition, it is possible to determine the metastatic potential of a tumor by determining the expression of a specific glycosyltransferase in the tumor cells. Useful for examining the expression of a specific glycosyltransferase gene are the Northern hybridization technique [Sambrook, Fritsch and Maniatis; Molecular Cloning—A Laboratory Manual, 2nd Edition, Cold Spring Harbor Laboratory Press, 1989], which uses, as a probe, the gene radioactively labeled, for example, and the polymerase chain reaction (hereinafter, “PCR”) technique [Innis et al.: PCR Protocols, Academic Press, 1990]. In applying these techniques, the specific glycosyltransferase gene or knowledge of the base sequence of the gene is required. From this viewpoint as well, it is important to clone the specific glycosyltransferase gene and determine its base sequence.
JP-A-2-227075 discloses the possibility of improving the properties of physiologically active proteins, such as granulocyte colony stimulating factor (G-CSF) and prourokinase (pro-UK), by artificially introducing a carbohydrate chain into the proteins using recombinant DNA technology.
As mentioned above, it is a very important problem from an industrial viewpoint to modify the structure of the carbohydrate chain of a glycoprotein or prepare a specific carbohydrate chain or a modification thereof in large quantities.
There have been marked advances in recent years in the means for modifying carbohydrate chain structures. In particular, it is now possible to structurally modify carbohydrate chains using highly specific enzymes (exoglucosidases) that are capable of releasing carbohydrate units one by one from the end of the carbohydrate chain, or glycopeptidases or endo-glycosidases that are capable of cleaving the site of binding to the peptide chain without causing any change in either the peptide or carbohydrate chains, and accordingly, to study biological roles of carbohydrate chains in detail. The recent discovery of endoglycoceramidases that are capable of cleaving the glycolipids at the site between the carbohydrate chain and the ceramide [Ito and Yamagata: Journal of Biological Chemistry, 262, 14278 (1986)] has not only made it easy to prepare carbohydrate chains of glycolipids but has also promoted investigations into functions of glycolipids, in particular glycolipids occurring in cell surface layers. Further, it has become possible to add new carbohydrate chains using glycosyltransferases. Thus, for instance, sialic acid can be added to a carbohydrate chain terminus using sialyltransferase [Sabesan and Paulson: Journal of the American Chemical Society, 108, 2068 (1986)]. It is also possible, using various glycosyltransferases or glycosidase inhibitors, to modify carbohydrate chains that are to be added [Allan et al.: Annual Review of Biochemistry, 56, 497 (1987)]. However, there is no means available for producing glycosyltransferases for use in synthesizing carbohydrate chains. It is desirable to produce glycosyltransferases in large quantities by cloning glycosyltransferases and causing efficient expression of glycosyltransferases in host cells utilizing recombinant DNA technology.
As regards α-1,3- or α-1,4-fucosyltransferase species possibly involved in the synthesis of the sialyl Lewis x or sialyl Lewis a carbohydrate chain, the occurrence of five enzyme activities has so far been suggested [Mollicone et al.: Carbohydrate Research, 228, 265-276 (1992); Weston et al. Journal of Biological Chemistry, 267, 24575-24584 (1992)]. Among genes coding for α-1,3- or α-1,4-fucosyltransferase, the following four have reportedly been isolated: α-1,3/1,4-fucosyltransferase gene (hereinafter referred to as “Fuc-TIII” for short) [Kukowska-Latallo et al.: Genes & Development, 4, 1288-1303 (1990)] directly involved in the synthesis of the Lewis blood type antigen carbohydrate chain; α-1,3-fucosyltransferase gene named EFLT (hereinafter, “Fuc-TIV”) [Goelz et al.: Cell, 63, 1349-1356 (1990)]; α-1,3-fucosyltransferase gene isolated by Weston et al. using the hybridization technique (hereinafter, “Fuc-TV”) [Weston et al.: Journal of Biological Chemistry, 267, 4152-4160 (1992)]; and α-1,3-fucosyltransferase gene isolated by Weston et al. using the hybridization technique (hereinafter, “Fuc-TVI”) [Weston et al.: Journal of Biological Chemistry, 267, 24575-24584 (1992)].
For the medical field, it is important to identify an α-1,3-fucosyltransferase that is directly involved in the synthesis of carbohydrate chains related to sialyl Lewis x, the ligand of ELAM-1 or GMP-140, in inflammatory leucocytes such as granulocytes.
Among the above-mentioned four α-1,3-fucosyltransferase genes Fuc-TIV by itself is apparently incapable of synthesizing the sialyl Lewis x carbohydrate chain, which is the ligand of ELAM-1, or the sialyl Lewis a carbohydrate chain [Lowe et al.: Journal of Biological Chemistry, 266, 17467-17477 (1991); Kumar et al.: Journal of Biological Chemistry, 266, 21777-21783 (1991)] and an α-1,3-fucosyltransferase directly involved in ELAM-1 ligand synthesis has not yet been obtained from human granulocytic or monocytic cells, for example the human granulocytic cell line HL-60 reportedly adhering to ELAM-1 [Lobb et al.: Journal of Immunology, 147, 124-129 (1991)].
In view of the foregoing, it is important to identify and isolate an α-1,3-fucosyltransferase that is directly involved in ELAM-1 ligand synthesis from human granulocytic or monocytic cells so that efficient in vitro or in vivo production of carbohydrate ligands directly involved in in vivo adhesion to ELAM-1 can be carried out. It is also important from the standpoint of detection of α-1,3-fucosyltransferase or inhibition of its production in sites of inflammation by the polymerase chain reaction technique utilizing DNA coding for the fucosyltransferase.
It is an object of the present invention to provide a novel α-1,3-fucosyltransferase species, a cDNA coding for the fucosyltransferase and a vector containing the cDNA, with which glycoproteins or glycolipids containing ligand carbohydrate chains of selectins, such as ELAM-1, can be efficiently produced in animal cells, in particular Namalwa cells. Another object is to provide a DNA coding for the fucosyltransferase useful in the treatment of diseases, such as inflammation, by inhibiting the expression of the fucosyltransferase using, for example, the above-mentioned antisense RNA/DNA technique or in the diagnosis of such diseases using, for example, the Northern hybridization or PCR technique.